The project goal was to build a converter that could translate the stats of entities from one legacy game system into the scheme of a later system. For each system, data for ~4,000 entities was cleaned and organized into a database. Percentile reference tables were generated for each attribute. The algorithm uses percentile mapping. Stats from the source game system are encoded into a percentile vector, then decoded into corresponding values in the target game system. Once that worked, other features were added, including database search and a summary technique for a hard-to-calculate stat.
Built with
- Machine Learning
- Data Analysis
- Python
- Data Wrangling
- PostgreSQL